Nondeterministic digital storytelling
In certain forms of writing, the idea of readers making choices about how to proceed is built right in. Think of textbooks — the chapters are generally modular, and it’s assumed that people may read them out of order, skip over some, or just pluck out one or two and ignore the rest.
Not so for novels and other narrative forms. Choice is a gimmick. The phrase “reader choice” conjures images of hokey Choose Your Own Adventure books, text adventures played at the command line on early personal computers, or, at best, the experimental works of the hypertext fiction movement of the 1980 and ’90s born of Hypercard and Eastgate Systems. To a mainstream audience, such artifacts, whatever we may think of as their artistic or historical significance, essentially scan as “weird stuff for computer nerds” or “games for children.” I think we can change this impression with a new approach to designing for reader choice in digital narrative.
To proceed into the Labyrinth of Impractical Ideas, Turn to Page 72.
In the landscape of storytelling media today, games are ascendant. The dominant mode of storytelling in gaming is immersion: the player controls a character who exists within a story world, with varying degrees of freedom to choose how to progress from one point to another. The player’s actions are constrained by the mechanics of the game, and the outcome of the game is dependent on the player’s choices. At a minimum, the player must progress through all areas of the game to access the full story, but at the other extreme the narrative will branch in different directions in response to the player’s choices.
Books and games, to state the obvious, tell stories in very different ways. But ebooks may have something to learn from games when it comes to the practice of design and the positioning of narratives as experiences rather than just stories.
Consider Super Mario Bros. 3, a classic title for the Nintendo Entertainment System. In this game, players traverse the game world with maps. Sometimes there are parts you don’t have to play — they can skip over them on their way to the ending, and the story, at least in terms of what is scripted, is not affected by that choice. The boss will always be the same, and the princess will always be in another castle (which is a whole other problem, but I digress). Even so, the player’s unique experience of playing through it — the events and the emotions she experiences in response to them — is what game designers call “emergent story.” It’s user generated, within the constraints set by the game designer, which means it’s never the same twice.
There’s not much story in Super Mario Bros. 3, so some more contemporary examples may shed more light on the challenges of designing for choice. Games have become more narratively complex and deploy increasingly sophisticated storytelling technologies: cinematic cutscenes, dialogue trees, environmental narrative, distributed collectibles with story content, even so-called “morality engines” that activate different paths through a branching narrative based on value judgments about the player’s actions. The proliferation of gaming and the democratization of game design technology has helped usher in the indie game movement, in which designers are increasingly pushing games outside the genres and conventions that have come to define commercial gaming. One trend in indie game design is games that are strictly narrative — they use game mechanics purely for content delivery.
Curtain is one compelling example. It’s a game about an abusive relationship. There are no points to win, no items to collect, no puzzles to solve. The primary game mechanic is to walk around an apartment and get emotionally abused. As you enter certain areas or inspect items in the environment, you trigger narration —what appear to be memories of toxic, destructive monologues from your partner. The goal of the game is to simulate and communicate a feeling, one where every single thing in your life triggers a painful memory, and there is no escape. A few key items trigger moves forward into a new area of the game, and it’s possible to go straight to them and speedrun the game in a minute or two, but what would be the point?
Gone Home is another story game, one that has won wide acclaim. The mechanics are similar in that you explore a house, gathering pieces of a story as you examine items, some of which convey information on their own and others that trigger nondiegetic audio diaries that more explicitly narrate a story. In this game, you are a college student in the 1990s who returns home from a semester abroad to find her family’s house empty and a cryptic note from her younger sister that suggests she’s run away or gone missing. Unlike in Curtain, all of the narrative content is backstory, most of it consisting of a series of letters your sister has written to you that gradually unfold the story of what happened to her and why she’s gone. This game is notable for the subtlety and emotional timbre of its writing and the immersiveness it achieves as it pushes gaming into genres where few others have dared to tread.
These are not adventure games or puzzle games or representatives of any other recognizable genre. For lack of a better name we might describe them as “literary games.” They are lofty and mature in their narrative ambitions, but they also embrace a less deterministic mode of storytelling that is familiar in gaming. To maintain immersion in a virtual world, game designers can encourage the player to access parts of a story in a certain order, but they often can’t strictly require that order without puncturing the illusion that enables suspension of disbelief. It creates too much separation between the character and the player if the character finds that all doors but one are locked in her own house, and each room contains the key to unlocking the next. The nature of this storytelling experience is fundamentally different than a book or movie. You, as the designer or “author” of the game, tell a story in a certain way, but you accept that there is a multiplicity of ways that people will experience it.
Which, of course, is hardly a new idea. The reader’s imagination has always been a key player in our understanding of literature.
“A book read by a thousand different people is a thousand different books.” — Andrei Tarkovsky
Still, you’ll find more and more examples of this sort of storytelling pattern when you look at How We Consume Digital Content Today™.
Twitter is a good example. People on Twitter often tell stories that are longer than a single tweet: sometimes these are middlebrow experiments such as a writer tweeting an entire short story one sentence at a time, but it happens much more frequently in everyday conversation. People have an argument to make or a story to tell and they just construct it in a series of tweets and accept that they can’t control how people will come to it. It may be through someone else’s reply or retweet, or it may be picked up in whole or part, in or out of order, in their followers’ timelines. The writer on Twitter is letting go of some control over her story. In return, we as readers may benefit from having more than one way of finding and experiencing the story.
Hopefully by now you are at least partially convinced that “nondeterministic digital storytelling,” a near-meaningless phrase that I just made up, is a thing. How can we turn it into a literary technique? This brings us to the question of how one designs for reader choice in a publishing system.
Compared to the stories I’ve described above, reading a novel involves very little choice. It’s a straight line. You start at Page 1 and keep turning to the next page until you get to the end. So what might an alternative model be? How about this: when you get to the end of a chapter, you have to do something. There is no default.
With that relatively simple change, we enter terra incognita for ebooks. EPUB and its related ebook specs require a default reading order. To bring about this new medium, we must imagine alternative ways of presenting content in a reading environment designed for choice.
Perhaps the most straightforward option is simply to display a menu of links at the end of each chapter. The reader chooses one, reads that chapter, chooses another one, and so on. Which links are available in which chapters is up to the author.
A further question raised by this idea is how the reader stays oriented in the story space and tracks her progress when the shape of it is unfamiliar or unknown. Ultimately this suggests that authors working in a nondeterministic storytelling medium need to conceive of their stories not just as collections of words but as user experiences to be designed. The medium affords opportunities not only for creative use of language but also for creative deployment and manipulation of the mechanisms used to deliver them.
As an example, let’s imagine a road novel: two teenage boys leaving home, driving from Brooklyn to California. It’s a familiar structure — they drive somewhere, something happens there, they move on to the next place. What if the author decides that readers should navigate the novel like his characters are navigating their journey — with a map? Because the story is about traveling, as we move through narrative time we also move through geographic space.
When we get to this point in the story, something new happens. Now the reader has a choice. Does she continue the forward progress of the journey, or does she go back to Brooklyn? What’s going on there? Is there another character there who we’re going to start following? Is it a flashback, moving the story backwards in time as well as space?
Let’s say she doesn’t want to interrupt her flow yet, so she presses forward.
Now more story is accumulating in Brooklyn, and the reader grows more curious. She is being pulled in two directions: whichever way she goes, she doesn’t know what she’s missed by not choosing the other path. As it happens, this mirrors the feelings that the main character in the novel is having: he is harboring second thoughts about leaving home, thinking about the people he’s left behind, wondering if his plans are going to work out.
As the writer in this scenario, I’ve given up some control over how my story gets read. In return, I’ve gained the ability to create a new kind of resonance between the plot and themes of my story and the experience of reading it.
We haven’t crossed fully over into the realm of games — the reader is not immersed in this world, and isn’t inhabiting a certain character, or any of the other familiar conventions of games that tell stories. But hopefully we’ve given the reader just enough of that exploratory freedom to create emotional effects that can be subtle — even, dare I say it, artful.
What we need to make this a reality is a scalable technical architecture for reading and writing systems that supports creative freedom in designing the user experience of story consumption while setting conceptual and practical boundaries that ensure its products are recognizable as “books,” rather than games or software applications or computer art.
So, let’s make that.
This essay is part of the (r)ebook project at the UC Berkeley School of Information. Thanks and credit to my research and design partners, Meredith Hitchcock and Paul Son, and to Jeremy Walsh for roughly equal measures of inspiration and desperation.